Evolving fuzzy neural networks for extracting rules

Two fuzzy neural network architectures are presented to realize knowledge extracting from input-output samples. The network parameters including the necessary membership functions of the input variables and the consequent parameters are tuned and identified using evolutionary programming. The trained networks are then pruned so that the general rules can be extracted and explained. The experimental results have shown that the similar classification rules can be obtained in comparison to that of other fuzzy neural approaches with less number of rules and membership functions